CN102262450B

CN102262450B - A kind of method and apparatus carrying out character conversion based on mixed defeated character string

Info

Publication number: CN102262450B
Application number: CN201010193167.6A
Authority: CN
Inventors: 张扬; 曹镇
Original assignee: Beijing Sogou Technology Development Co Ltd
Current assignee: Beijing Sogou Technology Development Co Ltd
Priority date: 2010-05-27
Filing date: 2010-05-27
Publication date: 2015-12-09
Anticipated expiration: 2030-05-27
Also published as: CN102262450A

Abstract

The invention provides a kind of method and apparatus carrying out character conversion based on mixed defeated character string, and a kind of character input method and input method system, wherein, described method of carrying out character conversion based on mixed defeated character string comprises: obtain mixed defeated character string; Distinguish mixed film trasport section dissimilar in mixed defeated character string; Based on all dissimilar mixed film trasport sections, and obtain this mixed transformation result of defeated character string under target language according to preset processing rule.The present invention when user adopts mixed defeated character string to input, also can obtain the input results needed for user.

Description

A kind of method and apparatus carrying out character conversion based on mixed defeated character string

Technical field

The present invention relates to input method technique field, particularly a kind of method and apparatus carrying out character conversion based on mixed defeated character string, and a kind of character input method and input method system.

Background technology

Chinese's ubiquity when remembering English word is difficult to spell, the problem that can not pronounce, such as, can pronounce but forget definite full spelling, or, know spelling but cacoepy really maybe can not pronounce; For the problems referred to above, existing input method system can provide some interpretative functions, and such as, providing with entry is granularity, and the service of word looked into by similar dictionary.

Such as, publication number is the Chinese patent open file of 200910039910.X, discloses a kind of technical scheme of character input method, and the translation steps of this technical scheme mainly completes in two steps: the first step, common Syllable text conversion; Second step, translates the candidate item that user chooses.This patent can input interpretative function according to user, but, need user to participate at translation process, affect Consumer's Experience; And described translation is mainly granularity with entry, and user is inputted to the translation process details of whole sentence, then do not describe in detail.

Further, existing various technical scheme does not consider that user carries out mixed defeated situation, and such as, when user inputs Sino-British mixed defeated character string " the special Buddhist nun of bri ", its transition translation both cannot be Chinese by prior art, also cannot be English by its transition translation.

In a word, the technical matters needing those skilled in the art urgently to solve is exactly: how can the translation translation function provided for mixed defeated character string of novelty.

Summary of the invention

Technical matters to be solved by this invention is to provide a kind of method and apparatus carrying out character conversion based on mixed defeated character string, can carry out character conversion based on the mixed defeated character string of user's input, obtains the transformation result (such as, translating) needed for user.

Accordingly, present invention also offers a kind of character input method and input method system, in order to carry out character conversion based on mixed defeated character string, and the representing of described character conversion result can be realized, user selects and exports, user can be realized and directly obtain screen in required transformation result when mixed defeated character string and export.

In order to solve the problem, the invention discloses a kind of method of carrying out character conversion based on mixed defeated character string, comprising: obtaining mixed defeated character string; Distinguish mixed film trasport section dissimilar in mixed defeated character string; Based on all dissimilar mixed film trasport sections, and obtain this mixed transformation result of defeated character string under target language according to preset processing rule.

Preferably, described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, and the step of described acquisition transformation result comprises: obtain the pronunciation unit sequence of pronunciation fragment under target language in described mixed defeated character string; Based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, obtain this mixed transformation result of defeated character string under target language.

Preferably, described dissimilar mixed film trasport section comprises the pronunciation fragment of target language and the pronunciation fragment of non-target language, and the step of described acquisition transformation result comprises: the pronunciation unit sequence of pronunciation fragment under target language obtaining the target language in described mixed defeated character string; Obtain the pronunciation unit sequence of pronunciation fragment under target language of the non-target language in described mixed defeated character string; Based on the pronunciation unit sequence under target language, obtain this mixed transformation result of defeated character string under target language.

Preferably, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising: obtain the pronunciation unit sequence of target language fragment under target language; Based on the pronunciation unit sequence of pronunciation fragment under target language and the pronunciation unit sequence of target language fragment under target language, obtain this mixed transformation result of defeated character string under target language.

Preferably, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising: based on the pronunciation unit sequence of pronunciation fragment under target language, obtain corresponding target language character string; Obtained target language character string is connected with described target language fragment, obtains this mixed transformation result of defeated character string under target language.

Preferably, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising: based on described target language fragment, in dictionary, matching inquiry obtains corresponding words set; Utilize the pronunciation unit sequence of pronunciation fragment under target language, in described words set, search obtains occurrence, as this mixed transformation result of defeated character string under target language.

Preferably, using the first fragment in described mixed defeated character string as pronunciation fragment, the second fragment, as target language fragment, obtains the transformation result under first object language; And using the second fragment in described mixed defeated character string as pronunciation fragment, the first fragment, as target fragment, obtains the transformation result under the second target language.

Preferably, using the pronunciation fragment of the first fragment in described mixed defeated character string as target language, the second fragment, as the pronunciation fragment of non-target language, obtains the transformation result under first object language; And using the pronunciation fragment of the second fragment in described mixed defeated character string as target language, the first fragment, as the pronunciation fragment of non-target language, obtains the transformation result under the second target language.

Preferably, can the information such as character types, upper and lower case letter, separator or state switching point be passed through, distinguish mixed film trasport section dissimilar in mixed defeated character string.

Preferably, described method can also comprise: the described mixed transformation result of defeated character string under target language is translated as another language words and shows.

Preferably, when user inputs long sentence or phrase, described method can also comprise: by the mode of separator or participle, obtain required mixed defeated character string the long sentence inputted or phrase from user.

Based on another embodiment of the present invention, also disclose a kind of device carrying out character conversion based on mixed defeated character string, comprising:

Interface unit, for obtaining mixed defeated character string;

Discrimination unit, for distinguishing mixed film trasport section dissimilar in mixed defeated character string;

Converting unit, for based on all dissimilar mixed film trasport sections, and obtains this mixed transformation result of defeated character string under target language according to preset processing rule.

Preferably, described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, and described converting unit comprises: the first syllable acquisition module, for obtaining the pronunciation unit sequence of pronunciation fragment under target language in described mixed defeated character string; First transformation result acquisition module, for based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, obtains this mixed transformation result of defeated character string under target language.

Preferably, described dissimilar mixed film trasport section comprises the pronunciation fragment of target language and the pronunciation fragment of non-target language, and described converting unit comprises:

Second syllable acquisition module, for obtaining the pronunciation unit sequence of pronunciation fragment under target language of the target language in described mixed defeated character string;

Triphone acquisition module, for obtaining the pronunciation unit sequence of pronunciation fragment under target language of the non-target language in described mixed defeated character string;

Second transformation result acquisition module, for based on the pronunciation unit sequence under target language, obtains this mixed transformation result of defeated character string under target language.

Preferably, described transformation result acquisition module specifically for: based on the pronunciation unit sequence of pronunciation fragment under target language and the pronunciation unit sequence of target language fragment under target language, obtain this mixed transformation result of defeated character string under target language.

Preferably, described transformation result acquisition module comprises:

Pronunciation fragments character obtains submodule, for based on the pronunciation unit sequence of pronunciation fragment under target language, obtains corresponding target language character string;

Splicing submodule, for obtained target language character string being connected with described target language fragment, obtains this mixed transformation result of defeated character string under target language.

Preferably, described transformation result acquisition module comprises:

Thesaurus-lookups submodule, for based on described target language fragment, in dictionary, matching inquiry obtains corresponding words set;

Syllable search submodule, for utilizing the pronunciation unit sequence of pronunciation fragment under target language, in described words set, search obtains occurrence, as this mixed transformation result of defeated character string under target language.

Preferably, described device can also comprise: Language Translation unit, for the described mixed transformation result of defeated character string under target language is translated as another language words.

Preferably, described device can also comprise: cutting unit, during for inputting long sentence or phrase as user, by the mode of separator or participle, obtains required mixed defeated character string the long sentence inputted or phrase from user.

Based on another embodiment of the present invention, also disclose a kind of character input method, comprising: the mixed defeated character string receiving user's input; Distinguish mixed film trasport section dissimilar in mixed defeated character string; Based on all dissimilar mixed film trasport sections, and obtain this mixed transformation result of defeated character string under target language according to preset processing rule, and shown; According to user's instruction, export the transformation result of specifying.

Preferably, using the first fragment in described mixed defeated character string as pronunciation fragment; Second fragment, as target fragment, obtains the transformation result under first object language; And using the second fragment in described mixed defeated character string as pronunciation fragment, the first fragment, as target fragment, obtains the transformation result under the second target language; Show the transformation result under two kinds of target languages obtained above.

Preferably, using the pronunciation fragment of the first fragment in described mixed defeated character string as target language, the second fragment, as the pronunciation fragment of non-target language, obtains the transformation result under first object language; And using the pronunciation fragment of the second fragment in described mixed defeated character string as target language, the first fragment, as the pronunciation fragment of non-target language, obtains the transformation result under the second target language; Show the transformation result under two kinds of target languages obtained above.

Preferably, described method also comprises: the described mixed transformation result of defeated character string under target language is translated as another language words and shows.

Preferably, when user inputs long sentence or phrase, described method also comprises: by the mode of separator or participle, obtain required mixed defeated character string the long sentence inputted or phrase from user.

According to another embodiment of the present invention, also disclose a kind of input method system, comprising:

Interface unit, for receiving the mixed defeated character string of user's input;

Converting unit, for based on all dissimilar mixed film trasport sections, and obtains this mixed transformation result of defeated character string under target language according to preset processing rule;

Display unit, for showing transformation result;

Output unit, for according to user's instruction, exports the transformation result of specifying.

Preferably, described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, and described converting unit comprises:

First syllable acquisition module, for obtaining the pronunciation unit sequence of pronunciation fragment under target language in described mixed defeated character string;

First transformation result acquisition module, for based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, obtains this mixed transformation result of defeated character string under target language.

Second transformation result acquisition module, for based on the pronunciation unit sequence of pronunciation fragment under target language, obtains this mixed transformation result of defeated character string under target language.

Preferably, described transformation result acquisition module comprises: pronunciation fragments character obtains submodule, for based on the pronunciation unit sequence of pronunciation fragment under target language, obtains corresponding target language character string; Splicing submodule, for obtained target language character string being connected with described target language fragment, obtains this mixed transformation result of defeated character string under target language.

Preferably, described transformation result acquisition module comprises:

Preferably, described input method system can also comprise:

Language Translation unit, for the described mixed transformation result of defeated character string under target language is translated as another language words, is shown by described display unit.

Preferably, described input method system can also comprise: cutting unit, during for inputting long sentence or phrase as user, by the mode of separator or participle, obtains required mixed defeated character string the long sentence inputted or phrase from user.

Compared with prior art, the present invention has the following advantages:

Forget the spelling of certain part of word user but remember that this part is pronounced, or user, when forgetting that the translation result list of certain part of word remembers that this part is spelt, just may use mixed defeated character string to represent the input words needed for oneself.

In the mixed defeated character string of user, pronunciation fragment usually can be adopted to represent forgotten part, and target language fragment represents other parts.For such characteristic, the present invention can simultaneously according to the pronunciation unit sequence of described pronunciation fragment under target language, and target language fragment, obtains this mixed defeated character string transformation result such as word spelling, complete translation result under target language.

Such as, Chinese user is when forgetting foreign language spelling, mixed defeated character string (a part of foreign language spelling, a part of foreign language pronunciation) just may be adopted to represent, then this mixed defeated character string just can be converted to the foreign language result needed for user by application the present invention.

Or when user forgets the spelling of whole word, macaronic pronunciation fragment also can be adopted to represent this word, also, described mixed defeated character string comprises macaronic pronunciation fragment; For this characteristic, first the present invention can obtain the pronunciation unit sequence of described macaronic pronunciation fragment under target language, then, based on this pronunciation unit sequence, obtain this mixed defeated character string transformation result such as word spelling, complete translation result under target language.

Accompanying drawing explanation

Fig. 1 is a kind of process flow diagram carrying out the embodiment of the method for character conversion based on mixed defeated character string of the present invention;

Fig. 2 is a kind of example of defeated character string mixed in long sentence being carried out to character conversion of the present invention;

Fig. 3 is a kind of structural drawing carrying out the device embodiment of character conversion based on mixed defeated character string of the present invention;

Fig. 4 is the process flow diagram of a kind of character input method embodiment of the present invention;

Fig. 5 be in the first input process of the present invention candidate item represent interface schematic diagram;

Fig. 6 be in the second input process of the present invention candidate item represent interface schematic diagram;

Fig. 7 be in the third input process of the present invention candidate item represent interface schematic diagram;

Fig. 8 be in the present invention's the 4th kind of input process candidate item represent interface schematic diagram;

Fig. 9 is the structural drawing of a kind of input method system embodiment of the present invention.

Embodiment

For enabling above-mentioned purpose of the present invention, feature and advantage become apparent more, and below in conjunction with the drawings and specific embodiments, the present invention is further detailed explanation.

Mixed defeated character string of the present invention can be any macaronic mixed defeated character string; Such as, Chinese and English, Chinese character and German, Japanese and English, Japanese and Arabic etc., below mainly mix defeated character string for China and Britain and be described, other mixed defeated character strings please refer to.It should be noted that, the present invention based on mixed defeated character string comprise pronunciation fragment, or, pronunciation fragment and target language fragment, wherein, in this mixed defeated character string, pronunciation fragment is used to pronouncing (target language and/or non-target language), another part is the fragment of target language, and based on this mixed defeated character string, the present invention can obtain the transformation result (target language words) under target language.

Various application scenarioss first possible to the present invention below simply illustrate, not as limit, just to better describing the present invention.

Application scenarios 1

User wishes to obtain foreign language spelling, does not but know the spelling of a wherein part, only remembers that foreign language pronounces, then apply the present invention, and the national word of transliteration or this country's pronunciation can be used to represent this part, thus the foreign language spelling needed for obtaining.

Such as, user is for the full name of input U.S. movie star Schwarzenegger, when being input to " schwar ", but find that part has not below been remembered clearly, therefore select to represent this part by this Chinese character " Singh " or Chinese pinyin " xin ' ge ", by the present invention according to described mixed defeated character string " schwar ' Singh " or " schwar ' xin ' ge ", obtain transformation result " Schwarzenegger ".

And for example, the present invention can obtain according to mixed defeated character string " this storehouse ool " " school ", obtain according to " playing city by you " " university ", obtain " miscellaneous " according to " misc Lei Niesi ", etc.

Application scenarios 2

User wishes the Chinese translation obtaining foreign language, does not but know the translation of a wherein part, only remembers English spelling, then apply the present invention, can use this part of foreign language character representation, thus the Chinese translation needed for obtaining.Such as, user is for knowing the Chinese translation of U.S. movie star Brittany (Brittney), but do not know the translation of front portion " bri ", therefore " the special Buddhist nun of bri ", as mixed defeated character string, is obtained transformation result " Bu Liteni " or " Brittney " by the present invention.

Application scenarios 3

User for knowing the foreign language spelling of China and foreign countries' compound word, then applies the present invention, can obtain the transformation result of the China and foreign countries such as " Karaoke ", " many A dreams " compound word: " Karaoke ", " Doraemon ".

With reference to Fig. 1, show a kind of process flow diagram carrying out the embodiment of the method for character conversion based on mixed defeated character string of the present invention, specifically can comprise:

Step 101, the mixed defeated character string of acquisition;

The present embodiment can be applied to various translation occasion, such as, can be the various softwares with interpretative function, or, the various websites etc. of translation service being provided, for carrying out character conversion based on mixed defeated character string, obtaining the described mixed transformation result of defeated character string under target language.

In practice, can input according to user and obtain described mixed defeated character string, such as, directly using user in the input " the special Buddhist nun of bri " of certain translation web site as mixed defeated character string; Or, can be selected by right button and obtain described mixed defeated character string, such as, can the interpretative function of certain software be integrated in right-click menu, like this, when user chooses " the special Buddhist nun of bri " character string and triggers the interpretative function of this right-click menu, can using " the special Buddhist nun of bri " the mixed defeated character string as the present embodiment.

Mixed film trasport section dissimilar in step 102, the mixed defeated character string of differentiation;

Mix defeated character string for China and Britain, its composition mainly contains following situation:

Situation 1, pronunciation fragment and target language fragment;

Generally speaking, Sino-British mixed defeated character string is Chinese character and English mixed defeated, now, only need determine the target language of the mixed defeated character string of these China and Britain, easily can mix acquisition defeated character string from these China and Britain and to pronounce fragment and target language fragment; Wherein, described target language can be specified by user, or, from the fragment of mixing two different languages that defeated character string comprises, determine that arbitrarily a language fragments is as target language.

For " the special Buddhist nun of bri ", include English fragment and Chinese fragment in " the special Buddhist nun of bri " character string, suppose that the target language that user specifies is English, so, namely target language fragment is wherein " bri ", and Chinese pronunciations fragment is then " Te Ni ".Or, when user does not specify, can determine that English is target language, also can determine that Chinese is target language; Using Chinese character as target language time, namely target language fragment is wherein " Te Ni ", and English equivalents fragment is then " bri "; Using English as target language time, namely target language fragment is wherein " bri ", and Chinese pronunciations fragment is then " Te Ni ".

But in some cases, Sino-British mixed defeated character string can also be Chinese pinyin and English mixed defeated, such as, user is for the full name of input U.S. movie star Schwarzenegger, when being input to " schwar ", but find that part has not below been remembered clearly, therefore select the Chinese pinyin " xinge " with this part " Singh ".

Now, described mixed defeated character string can adopt separator " ' " connect described Chinese pinyin and English, such as.

Now, by the content of Chinese pronunciations fragment " xinge " as described in distinguishing as under type and English fragment " schwar ":

Such as, using capitalization as English fragment, non-capital character is as Chinese pronunciations fragment, and now, described mixed defeated string can be " SCHWARxinge "; Or switch the word of English input as English fragment by shift, the word inputted under input in Chinese pattern is as Chinese pronunciations fragment; Or, first utilize special symbol by described English fragment and Chinese pronunciations fragment separated, then, determine the language belonging to described two fragments; Such as, for " schwar ' xinge " mixed defeated string, can first utilize separator " ' " described two fragments are made a distinction.

Because Chinese pinyin and English are letter spelling, in practice, according to spelling conventions, the language belonging to a certain fragment can be determined.Such as, Chinese character has 415 basic syllables (not containing the syllable of tone), can learn that " xinge " belongs to Chinese pronunciations fragment accordingly; And for example, English word has fixing spelling conventions, and such as alphabetical fragment " qw ", " qd ", " bd ", " ooo " are illegal, and " ps ", " tr " can appear at word beginning, can obtain " schwar " accordingly and belong to English.

Like this, for the fragment of a mixed defeated character string, can according to the situation of the spelling conventions of described English word and Chinese syllable, the fragment computations that this mixed defeated character string is comprised its belong to the probability of English word and it belongs to the probability of Chinese syllable, to determine that certain alphabetical fragment is wherein Chinese pinyin or English.

Be appreciated that described pronunciation fragment can be any one in Chinese pronunciations fragment, English equivalents fragment, or the two combination, even also can represent for the pronunciation of other types, the present invention is not limited this.

Such as, the teacher==powerful army of v5, fragment of wherein pronouncing is made up of the pronunciation of English alphabet " V " and the Chinese pronunciations of numeric character " 5 ".

Again such as, 8 mistakes==good, wherein, pronunciation fragment is exactly numeric character " 8 ", but its formal Chinese pronunciations is " ba ", instead of " bu ", and can be identified as " bu " in the present invention.Namely, in specific implementation of the present invention, according to user's use habit, the mapping that numeric character " 8 " arrives " ba " and " bu " can be formed; Do not limit to the pronunciation standard formal with it, can be out of shape.

The pronunciation fragment of situation 2, target language and the pronunciation fragment of non-target language.

Such as, mix defeated string " Karaoke " for China and Britain, " OK a karaoke club " wherein can be divided into Chinese pronunciations fragment, " OK " divides into English equivalents fragment.

Step 103, based on all dissimilar mixed film trasport sections, and obtain this mixed transformation result of defeated character string under target language according to preset processing rule.

Such as, when described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, described step 103 can comprise:

Sub-step P1, the pronunciation unit sequence of pronunciation fragment under target language obtained in described mixed defeated character string;

Sub-step P2, based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, obtain this mixed transformation result of defeated character string under target language.

And for example, when described dissimilar mixed film trasport section comprises the pronunciation fragment of the pronunciation fragment of target language and non-target language, described step 103 can comprise:

Sub-step Q1, obtain the pronunciation unit sequence of pronunciation fragment under target language of the target language in described mixed defeated character string;

Sub-step Q2, obtain the pronunciation unit sequence of pronunciation fragment under target language of the non-target language in described mixed defeated character string;

Sub-step Q3, based on the pronunciation unit sequence under target language, obtain this mixed transformation result of defeated character string under target language.

Wherein, described pronunciation unit sequence can be understood as the combination of the pronunciation unit under certain language.

Such as, for Chinese, described pronunciation unit is exactly pinyin syllable (e.g. ,/chen/ ,/ya/ ,/da/ etc.), and like this, namely the pronunciation unit sequence of English equivalents fragment under Chinese be Chinese pinyin.

For English, described pronunciation unit is exactly English phonetic symbol (e.g., vowel/i/ ,/ai/, consonant/b/ ,/ts/ etc.), and like this, namely the pronunciation unit sequence of Chinese pronunciations fragment under English be English phonetic symbol.

In specific implementation, by enunciative mapping relations between bilingual, can adopt and obtain the pronunciation unit sequence of various pronunciation fragment under target language in various manners.

Such as, for Chinese character and English mixed defeated character string, if target language is English, so, the process obtaining the pronunciation unit sequence of this Chinese character under English can comprise: Chinese character → Chinese pinyin → English phonetic symbol, and wherein, " → " represents conversion or mapping relations.

For Chinese pinyin and English mixed defeated character string, if target language is English, so, the process obtaining the pronunciation unit sequence of this Chinese pinyin under English can comprise: Chinese pinyin → English phonetic symbol.

If target language is Chinese, then the process obtaining the pronunciation unit sequence of English fragment under Chinese can comprise: English → English phonetic symbol → Chinese pinyin.

Again such as, for mixed characters string that the is English and language such as German or French; English fragment → English phonetic symbol → German or French phonetic symbol.

The implementation of following sub-paragraphs P1-P2 is described in detail, and it is for mixed defeated character string, simultaneously according to pronunciation fragment wherein and target language fragment, can obtain this mixed transformation result of defeated character string under target language.Specifically, be first the pronunciation unit sequence under target language by pronunciation fragment map wherein, then, utilize the contact of described pronunciation unit sequence and target language fragment, obtain this mixed transformation result of defeated character string under target language.

Such as, mix defeated character string for China and Britain, sub-step P2 can be according to the Chinese pinyin of English equivalents fragment under Chinese, and Chinese fragment (Chinese character or Chinese pinyin), obtains Chinese character; Or, according to the English phonetic symbol of Chinese pronunciations fragment under English, and English fragment, obtain English word.

The present invention can provide following sub-step P2 to obtain the scheme of transformation result:

Scheme one,

The thought of this programme is, utilizes pronunciation fragment and target language fragment in enunciative contact, integrates the pronunciation unit sequence of described two fragments; Now, described sub-step P2 can comprise:

Sub-step A1, the pronunciation unit sequence of acquisition target language fragment under target language;

Sub-step A2, based on the pronunciation unit sequence of pronunciation fragment under target language and the pronunciation unit sequence of target language fragment under target language, obtain this mixed transformation result of defeated character string under target language.

For " the special Buddhist nun of bri ", suppose that target language fragment is wherein " bri ", pronunciation fragment " Te Ni ".Then by sub-step P2, obtain fragment " Te Ni " the pronunciation unit sequence under target language (English), also namely English phonetic symbol is /tni/, so, then can be obtained the English phonetic symbol/bri/ of " bri " by sub-step A1.And first sub-step A2 can carry out possible integration to described two English phonetic symbols, such as one of them integrated results is [britni], then, according to this integrated results [britni], mapped by modes such as retrieval dictionaries and obtain transformation result " Britney ".

In addition, all right other English phonetic symbol corresponding of the pronunciation fragment " Te Ni " in above-mentioned example, as like this, can comprise in the integrated results that sub-step A2 obtains be " Brittany " by the retrieval mode such as dictionary and then the transformation result that obtains.

In a word, this programme obtains this mixed pronunciation unit sequence of defeated character string under target language by integrating, and then maps and obtains corresponding transformation result.

Scheme two,

The thought of this programme is, utilizes pronunciation fragment and the contact of target language fragment on character, connects the character string of described two fragments; Now, described sub-step P2 can comprise:

Sub-step B1, based on the pronunciation pronunciation unit sequence of fragment under target language, obtain corresponding target language character string;

Sub-step B2, obtained target language character string to be connected with described target language fragment, to obtain this mixed transformation result of defeated character string under target language.

For " bri special Buddhist nun ", suppose that target language fragment is wherein " bri ", then by sub-step P1, the English phonetic symbol obtaining fragment " Te Ni " correspondence is /tni/ or so, then/English character the sequence of tni/ can be obtained by sub-step B1: teny, or, english character sequence: taney.

Like this, target language fragment is that " bri " is connected with the English character sequence that sub-step B1 obtains by sub-step B2, and the transformation result obtained can be " Briteny ", or, " Brittany ".

For the situation that target language is Chinese, Chinese fragment in " the special Buddhist nun of bri " is " Te Ni ", suppose that it passes through step 102, the Chinese pinyin of pronunciation fragment " bri " correspondence is /buli/, so, sub-step B1 can obtain/sequence of Chinese characters of buli/, also be Chinese character, be assumed to be " in cloth ", so, described " in cloth " can be connected with " Te Ni " by sub-step B2, and obtaining transformation result can be " Bu Liteni ".

No matter target language is English or Chinese, character can as the basic composition unit of this mixed transformation result of defeated character string under target language, further, target language fragment again can as a kind of character string in this transformation result, like this, only need obtain the target language character string of pronunciation field, then connect described two kinds of character strings, described transformation result can be obtained.

Scheme three,

Above-mentioned two schemes utilize the thought of part → integration, by setting up pronunciation fragment and the contact of target language fragment in pronunciation or character, obtain described transformation result.Provide another implementation of the present invention below:

Described sub-step P2 can realize this programme by following sub-step:

Sub-step C1, based on described target language fragment, in dictionary, matching inquiry obtains corresponding words set;

Sub-step C2, the pronunciation unit sequence of utilization pronunciation fragment under target language, in described words set, search obtains occurrence, as this mixed transformation result of defeated character string under target language.

For " chau not ", suppose that target language fragment is wherein " chau ", then the English phonetic symbol obtaining fragment " not " corresponding by sub-step P1 is so, in dictionary, the set of letters (such as, now pronunciation fragment can be substituted in the mode of asterisk wildcard) started with " chau " can be obtained by matching inquiry by sub-step C1, such as { chauffeur, chauvinism ..., chauvinist}; Sub-step C2 then can utilize phonetic symbol being search occurrence chauffeur in described set of letters, is also transformation result of the present invention.

For mixed defeated character string " chau not ", require that its transformation result under target language is with target language fragment " chau " beginning, character string is below close with the pronunciation of pronunciation fragment " not " simultaneously; Because target language fragment is as the composition character string of transformation result, and fragment of pronouncing can limit the pronunciation of other parts except target language fragment in transformation result, therefore, this programme can realize a kind of words coupling of limited range, reduce matching range, promote the efficiency of mixed defeated character string to transformation result.

Be appreciated that the dictionary of this programme can replace by the data structure that other has words memory function, the present invention does not need to be limited this; Such as, described data structure can be positioned on Local or Remote server.

Be described in detail several schemes obtaining transformation result above, described several scheme for the mixed defeated character string under several scenes, can obtain this mixed transformation result of defeated character string under target language, can meet consumers' demand; Be appreciated that those skilled in the art can several scheme described in conbined usage as required, or use wherein any one scheme, the present invention is not limited to this.

For making those skilled in the art understand the present invention better, providing the specific embodiment that the transfer process of defeated character string is mixed by China and Britain below, can comprise:

Step S1, the Sino-British mixed defeated character string of acquisition; Defeated character string is mixed by described China and Britain can comprise pronunciation fragment and target language fragment;

Step S2, obtain the pronunciation unit sequence of described pronunciation fragment under target language, specifically can comprise:

Sub-step S21, for described pronunciation fragment, cutting obtains corresponding source language syllable unit;

This step is by fragment of pronouncing, and is cut into fine-grained source language syllable unit s1s2s3...sn.

For English equivalents fragment, cutting is according to being consonant driving, substantially be that a syllable is (if vowel before current consonant (combination) to next consonant, then have a syllable of one's own), but need to consider the situation of consonant liaison (two n of writing, two write l etc.) or special pronunciation combination (such as "-er ", "-er-", " wh-").

Such as, 3 English syllable unit can be obtained for " schwar " cutting: sch, wa and r; 2 English syllable unit: b and ri are obtained for bri cutting.

For Chinese pronunciations fragment, if Chinese character, then first Chinese character being converted into Chinese pinyin, as " Te Ni " → " teni ", then, is pinyin syllable unit by this Chinese pinyin cutting, and such as, the cutting result of " teni " can be: te and ni.

Certainly, for the present invention, if described pronunciation fragment comprises multiple syllable unit, then can to its cutting, if described pronunciation fragment only comprises a syllable unit, then the present invention also can not carry out dicing step, this syllable unit of Direct Recognition.Such as, " chau " only uses a consonant, therefore only comprises a syllable unit.

Sub-step S22, described source language syllable unit is converted into source language pronunciation unit;

Described conversion can be the mapping of a multi-to-multi.Table 1 gives the mapping table example of English syllable unit to English phonetic symbol unit, and wherein "-" represents the front and back position of syllable in word; Like this, by look-up table 1, corresponding source language pronunciation unit can be found for the source language syllable unit in pronunciation fragment, such as, for " schwar ", described in the process of consulting can be: sch →/∫/, war →/va/.

Table 1

In some embodiments of the invention, the mapping of source language syllable unit to target language pronunciation unit can directly be realized.Such as, for Chinese, pinyin syllable unit inherently pronounces unit, therefore without the need to the conversion operation of this step.

But for English, because some English syllable unit exists multiple pronunciation situation, diverse location or context difference is in entry, the sound then sending out different, therefore, in order to improve the accuracy of transliteration, first English syllable unit is converted to English phonetic symbol unit, because the pronunciation of English phonetic symbol unit has uniqueness.

Namely in this example, pronounce from source language syllable unit to target language the conversion of unit, not only directly change according to source language syllable unit, but also with reference to the position of source language syllable unit in pronunciation fragment, and/or, its contextual situation, by the pronunciation unit of the syllable unit converting into target language of source language.

Sub-step S23, be target language pronunciation unit by described source language pronunciation unit maps;

Table 2 is examples that English phonetic symbol arrives Chinese pinyin.This can be the mapping of a multi-to-multi equally, and this has just constructed described pronunciation fragment, to be pronounced the pronunciation unit networks that unit forms by target language.Like this, by look-up table 2, mapping relations can be found for the English phonetic symbol in pronunciation fragment " schwar " :/∫/→ shi/si ,/va/ → wa.

Table 2

English phonetic symbol	Chinese pinyin
		/ts/	ci
/b/	bu
		/va/	wa
/ri/	li
		/∫/	shi/si
/v/，/vu/	fu，wu
		/wu/	wu
......	......

In theory, for any bilingual (as, Chinese is with English, Japanese and Arabic etc.), the mapping relations of pronouncing on unit separately can be based upon, thus obtain source language syllable unit after cutting can being carried out to pronunciation fragment, the pronunciation unit of target language is then obtained based on the enunciative association of bilingual.

Concrete, the source language related in the present embodiment and target language can be double-byte characters word (Chinese written language, Japanese character etc.) and single-byte character word (English etc.).Further, for double-byte characters word, described pronunciation fragment can comprise shape code (Chinese five, Korean) character string or tone code (Chinese pinyin, Japanese) character string.

In the above example, English syllable unit and English phonetic symbol unit is introduced as intermediary in the transfer process of step S2 English equivalents fragment → Chinese pinyin unit, namely have employed the process of English equivalents fragment → English syllable unit → English phonetic symbol unit → Chinese pinyin unit, to improve the accuracy of English syllable sounds, improve the order of accuarcy of conversion.

In a word, this step is exactly the mapping relations needing to find pronunciation fragment → target language pronunciation unit sequence, in order to accuracy or other better effects, can introduce intermediary, adopt the implementation of indirect conversion.

Step S3, the pronunciation unit sequence of acquisition target language fragment under target language, specifically can comprise:

Such as, for target language fragment (Chinese fragment) " Te Ni ", Chinese pinyin " teni " first can be translated into, then, be pinyin syllable unit: te and ni by this Chinese pinyin " teni " cutting, the pronunciation unit of described pinyin syllable unit also i.e. Chinese.

Due to its implementation procedure and aforementioned exemplary similar, relevant part is see illustrating, therefore not to repeat here.

Step S4, based on the pronunciation unit sequence of pronunciation fragment under target language and the pronunciation unit sequence of target language fragment under target language, obtain this mixed transformation result of defeated character string under target language.

In specific implementation, to each target language pronunciation unit generated in step S2 and S3, by look-up table 3 (for translated name, the example of a translated name word table), for its generate target word set di1, di2 ... dini}.Such as, the 1st target language pronunciation unit has n ₁individual target word candidate, second target language pronouncing unit has n ₂individual target word candidate, the n-th target language pronunciation unit has n _nindividual transformation result candidate.The candidate one of so corresponding with whole mixed defeated character string transformation result has n ₁× n ₂× ... n _n=m, which forms a candidate conversion result network.

Table 3

Chinese pinyin syllable	Wait selected character
		bu	Cloth, or not step
li	In, from
		te	Special
ni	Girl, you, Buddhist nun
		si	Four, think, seemingly
shi	Be, city, execute
		Wa	Baby, watt
......	......

Then, in this transformation result candidate network, optimum/n-best path is found.In these all m candidates, calculate the language model scores generating candidate, find out optimum 1 or n bar optimal path, export as transliteration translation result.The method finding optimum/n-best path in concrete network is applied in the prior art to some extent, such as, and the long sentence input of existing Chinese pinyin input method.Provide an example of the present invention to be below briefly described.

For the mixed defeated character string " the special Buddhist nun of bri " that target language is Chinese, suppose that the Chinese pronunciations unit that step S2 acquires has: bu, li, the Chinese pronunciations unit that step S3 acquires has te, ni, and this step is then for described 4 Chinese pronunciations unit generate target word; By look-up table 3, bu{ cloth, or not step ..., in li{, from ..., te{ is special ..., ni{ girl, you, Buddhist nun ..., then the candidate one of the transformation result that defeated character string mixed with this is corresponding has 3 × 2 × 1 × 3=18; , T.G Grammar result candidate network; The language model scores of last calculated candidate, obtain two the highest candidates's " Brittney " of score and " Bu Liteni ", alternatively item provides.

Language model is wherein some characterization rules provided based on speech habits and usage, concrete, can based on the statistics of existing translated name (as, use word frequency, where be commonly used in: translated name beginning/centre/end etc.), object is the custom making the translated name result automatically generated more meet artificial translated name.Such as, such as " baby ", is generally used in woman's name end; And " watt " of unisonance, substantially can be used in the optional position of translated name.The present invention does not need to be limited for the setting of concrete syntax model, and those skilled in the art carry out setting according to the characteristic of concrete syntax, use habit and the syntax etc.

In addition, in the above example, the English mapping relations such as pronunciation mapping table, translated name word table to Chinese character can be that manual sorting obtains, and also can be to obtain based on the programming count of language material.The present invention does not need to be limited concrete generative process.

In one preferred embodiment of the invention, according to user's request, the described mixed transformation result of defeated character string under two kinds of target languages can be provided.

One, when described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, the present embodiment can comprise:

Using the first fragment in described mixed defeated character string as pronunciation fragment, the second fragment, as target fragment, obtains the transformation result under first object language;

And using the second fragment in described mixed defeated character string as pronunciation fragment, the first fragment, as target fragment, obtains the transformation result under the second target language.

Such as, for mixed defeated character string " the special Buddhist nun of bri ", the present embodiment can provide the transformation result " Brittney " under Chinese, and the transformation result " Britney " under English.In practice, above-mentioned functions can the form of function button be presented in translation web site, or as the function i ntegration of translation software at right-click menu, the present invention is not limited this.

Two, when described dissimilar mixed film trasport section comprises the pronunciation fragment of the pronunciation fragment of target language and non-target language, the present embodiment can comprise:

Using the pronunciation fragment of the first fragment in described mixed defeated character string as target language, the second fragment, as the pronunciation fragment of non-target language, obtains the transformation result under first object language;

And using the pronunciation fragment of the second fragment in described mixed defeated character string as target language, the first fragment, as the pronunciation fragment of non-target language, obtains the transformation result under the second target language.

In another preferred embodiment of the invention, obtain this mixed defeated character string after the transformation result under target language in step 103, this transformation result described can also be translated as another language words and show by the present invention.Such as, for mixed defeated character string " the special Buddhist nun of bri ", after the transformation result " Brittney " under step 103 obtains Chinese, translation can also be carried out to " Brittney " and obtain Brittany, then show described two results simultaneously.

It should be noted that, the present invention also may be used for the entirety translation of some long sentences or phrase, now, can pass through the mode of separator or participle, obtain required mixed defeated character string the long sentence inputted or phrase from user.

Usually, the word in long sentence is separated in user habit space, with reference to Fig. 2, in a kind of application example of the present invention, user view input " Thechaufferisonthephone ", but because the latter half of word " chauffer " is difficult to memory, so user replaces with Chinese character " not ", also namely, the long sentence of user's input is " Thechau is isonthephone not "; Therefore, the present invention can obtain required mixed defeated character string " chau not " from this long sentence, and after obtaining the transformation result " chauffeur " of " chau not ", shows described long sentence with the form of candidate item.

When user does not adopt space to separate the word in long sentence, the present invention also can obtain wherein required mixed defeated character string by the mode of participle, certainly, and the situation that its accuracy may be separated less than user oneself.

Further, when target language fragment exists mistake, the present invention can also do correction process; Such as, the transformation result " Brittney " under Chinese that the present invention obtains " the special Buddhist nun of bri " is namely be corrected as " Te Ni " Chinese fragment " Te Ni ".

Moreover consider the upper and lower case letter transfer problem often run in English, the present embodiment can support specific capital and small letter matched rule, can provide the transformation result of applicable demand; Such as, automatically by the initial caps of first word in long sentence, or automatically by the initial caps of the word such as name, place name, the transformation result as " the special Buddhist nun of bri " can comprise " Brittany ", etc.

Above macaronic mixed defeated character string is introduced, it should be noted that, in described mixed defeated character string, the mixed defeated of numeral, character and language can also be comprised, as long as fragment of namely pronouncing can embody, certain specific language might not be belonged to.Such as, based on mixed defeated character string " 7 tea ", its pronunciation fragment is " 7 ", and target language fragment is " tea ", obtains transformation result and " infuses (qi) tea ".And for example, based on mixed defeated character string "+people ", its pronunciation fragment is "+", and target language fragment obtains transformation result " household " for " people ", etc.

Corresponding with preceding method embodiment, the present invention gives a kind of device carrying out character conversion based on mixed defeated character string, and with reference to the structural drawing shown in Fig. 3, described device specifically can comprise:

Interface unit 301, for obtaining mixed defeated character string; Discrimination unit 302, for distinguishing mixed film trasport section dissimilar in mixed defeated character string;

Converting unit 303, for based on all dissimilar mixed film trasport sections, and obtains this mixed transformation result of defeated character string under target language according to preset processing rule.

In practice, according to the mixed defeated character string of different situation, described converting unit 303 can be designed:

Situation 1, described dissimilar mixed film trasport section comprise pronunciation fragment and target language fragment time, the structure of described converting unit 303 can comprise:

This situation, for mixed defeated character string, simultaneously according to pronunciation fragment wherein and target language fragment, can obtain this mixed transformation result of defeated character string under target language.In specific implementation, the structural design scheme of following transformation result acquisition module can be provided:

Design proposal one,

This programme utilizes pronunciation fragment and target language fragment in enunciative contact, integrates the pronunciation unit sequence of described two fragments; Now, described transformation result acquisition module specifically may be used for:

Based on the pronunciation unit sequence of pronunciation fragment under target language and the pronunciation unit sequence of target language fragment under target language, obtain this mixed transformation result of defeated character string under target language, and shown.

Design proposal two,

This programme utilizes pronunciation fragment and the contact of target language fragment on character, connects the character string of described two fragments; Based on this thought, can design as lower module in described transformation result acquisition module:

Pronunciation fragments character obtains submodule D1, for based on the pronunciation unit sequence of pronunciation fragment under target language, obtains corresponding target language character string;

Splicing submodule D2, for obtained target language character string being connected with described target language fragment, obtaining this mixed transformation result of defeated character string under target language, and being shown.

Design proposal three,

In the program, described transformation result acquisition module can comprise as lower module:

Thesaurus-lookups submodule E1, for based on described target language fragment, in dictionary, matching inquiry obtains corresponding words set;

Syllable search submodule E2, for utilizing the pronunciation unit sequence of pronunciation fragment under target language, in described words set, search obtains occurrence, as this mixed transformation result of defeated character string under target language.

Be appreciated that one or more that those skilled in the art can use in described several scheme as required, the present invention is not limited this.

Situation 2, when described dissimilar mixed film trasport section comprises the pronunciation fragment of the pronunciation fragment of target language and non-target language, described converting unit 303 may further include:

Can there is similarity in some English word, the English phonetic symbol as movie star's Blanc girl " Britney " is [britni], and the English phonetic symbol of Brittney " Brittany " is in spelling and pronunciation like this, if user is for obtaining the English full name " Britney " of Blanc girl by mixed defeated character string " the special Buddhist nun of bri ", but when providing two transformation results " Britney " and " Brittany ", user does not but know which is the word wanted.

For said circumstances, in one preferred embodiment of the invention, described device can also comprise:

Language Translation unit, for being translated as another language words by the described mixed transformation result of defeated character string under target language.Such as, above-mentioned two transformation results " Britney " and " Brittany " are translated as " Blanc girl " and " Brittney ".

The preferred embodiment can make user according to described translation result, selects corresponding transformation result, improves the accuracy of conversion.

In addition, the present invention according to user's request, can provide the described mixed transformation result of defeated character string under two kinds of target languages:

Such as, when described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, can design as lower module in described converting unit 303:

First modular converter F1, for using the first fragment in described mixed defeated character string as pronunciation fragment, the second fragment, as target fragment, obtains the transformation result under first object language;

Second modular converter F2, for using the second fragment in described mixed defeated character string as pronunciation fragment, the first fragment, as target fragment, obtains the transformation result under the second target language

And for example, when described dissimilar mixed film trasport section comprises the pronunciation fragment of the pronunciation fragment of target language and non-target language, can design as lower module in described converting unit 303:

3rd modular converter, for mixing the pronunciation fragment of the first fragment in defeated character string as target language using described, the second fragment, as the pronunciation fragment of non-target language, obtains the transformation result under first object language;

4th modular converter, using the pronunciation fragment of the second fragment in described mixed defeated character string as target language, the first fragment, as the pronunciation fragment of non-target language, obtains the transformation result under the second target language.

It should be noted that, the present invention also may be used for the entirety translation of some long sentences or phrase, now, the custom of the separators such as space is used for user, can design in said device: cutting unit, during for inputting long sentence or phrase as user, by the mode of separator or participle, the long sentence inputted from user or phrase, obtain required mixed defeated character string.

With reference to Fig. 4, show the process flow diagram of a kind of character input method embodiment of the present invention, specifically can comprise:

The mixed defeated character string of step 401, reception user input;

Mixed film trasport section dissimilar in step 402, the mixed defeated character string of differentiation;

Step 403, based on all dissimilar mixed film trasport sections, and obtain this mixed transformation result of defeated character string under target language according to preset processing rule, and shown;

Step 404, foundation user instruction, export the transformation result of specifying.

The present embodiment can be applied in the input process of various word, for the mixed defeated character string inputted according to user, show the one or more transformation results obtained, wherein, described transformation result can the form of candidate item, feedback display is selected to user, then exports the transformation result that user selects.Be about to the translation conversion plan that the present invention is based on mixed defeated character string, be applied to concrete character input method field.

With reference to Fig. 5, user is for by the mixed defeated character string of input " schwar ' xin ' ge ", and to obtain the Full Name in English of U.S. movie star Schwarzenegger, the present invention according to described mixed defeated character string, can obtain transformation result candidate item: 1, Schwarzenegger; Like this, user can select this candidate item is shielded as required.

With reference to Fig. 6, user, for by mixed defeated character string " bri spy Buddhist nun ", obtains the Chinese translation of U.S. movie star Brittany (Brittney), and the present invention is according to described mixed defeated character string, and the candidate item obtained can comprise: 1, Brittney; 2, Bu Litani; 3, Blanc girl; Like this, user knocks in numeral 1,2 or 3, can complete the upper screen of corresponding candidate item.

With reference to Fig. 7, show the input example that the present invention's a kind of day English mixes defeated string, user wants the Japanese word " メモリ " inputting " internal memory ", when being input to the first two character, but the 3rd alphabetical spelling is forgotten, so the English character " ry " of input unisonance; Like this, the present invention according to mixed defeated character string " メモ ry ", can obtain transformation result candidate item: " メモリ ", is selected as required this candidate item is shielded by user.

Such as, when described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, simultaneously according to pronunciation fragment wherein and target language fragment, this mixed transformation result of defeated character string under target language can be obtained; Now, described step 403 can comprise:

Sub-step R1, the pronunciation unit sequence of pronunciation fragment under target language obtained in described mixed defeated character string;

Sub-step R2, based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, obtain this mixed transformation result of defeated character string under target language.

For sub-step R2, the present embodiment can provide following scheme:

Scheme one,

This programme utilizes pronunciation fragment and target language fragment in enunciative contact, integrates the pronunciation unit sequence of described two fragments; Now, described sub-step R2 can comprise:

Sub-step G1, the pronunciation unit sequence of acquisition target language fragment under target language;

Sub-step G2, based on the pronunciation unit sequence of pronunciation fragment under target language and the pronunciation unit sequence of target language fragment under target language, obtain this mixed transformation result of defeated character string under target language.

Scheme two,

This programme utilizes pronunciation fragment and the contact of target language fragment on character, connects the character string of described two fragments; Based on this thought, described sub-step R2 can comprise:

Sub-step H1, based on the pronunciation pronunciation unit sequence of fragment under target language, obtain corresponding target language character string;

Sub-step H2, obtained target language character string to be connected with described target language fragment, to obtain this mixed transformation result of defeated character string under target language.

Scheme three,

In the program, described sub-step R2 can comprise following sub-step:

Sub-step I1, based on described target language fragment, in dictionary, matching inquiry obtains corresponding words set;

Sub-step I2, the pronunciation unit sequence of utilization pronunciation fragment under target language, in described words set, search obtains occurrence, as this mixed transformation result of defeated character string under target language.

In one preferred embodiment of the invention, according to user's request, described mixed defeated character can be provided

The transformation result of string under two kinds of target languages, now, described character input method specifically can comprise:

And for example, when described dissimilar mixed film trasport section comprises the pronunciation fragment of the pronunciation fragment of target language and non-target language, described step 403 may further include:

Obtain the pronunciation unit sequence of pronunciation fragment under target language of the target language in described mixed defeated character string;

Obtain the pronunciation unit sequence of pronunciation fragment under target language of the non-target language in described mixed defeated character string;

Based on the pronunciation unit sequence of pronunciation fragment under target language, obtain this mixed transformation result of defeated character string under target language.

Under this kind of situation, for providing the described mixed transformation result of defeated character string under two kinds of target languages, described character input method specifically can comprise:

And using the pronunciation fragment of the second fragment in described mixed defeated character string as target language, the first fragment, as the pronunciation fragment of non-target language, obtains the transformation result under the second target language;

Show the transformation result under two kinds of target languages obtained above.

In addition, select the candidate item wanted for the convenience of the user, and improve the accuracy rate of input, in another preferred embodiment of the invention, obtain this mixed defeated character string after the transformation result under target language in step 403, this transformation result described can also be translated as another language words by the present invention.

Such as, if user is for obtaining the English full name " Britney " of Blanc girl by mixed defeated character string " the special Buddhist nun of bri ", step 403 acquires two transformation results: time " Britney " and " Brittany ", and user does not but know which is the word wanted.

This preferred embodiment for " Britney " and " Brittany ", obtains translation result respectively: " Blanc girl " and " Brittney "; With reference to Fig. 8, the candidate item that the present invention obtains can comprise: 1, Britney; 2, Blanc girl; 3, Brittany; 4, Bu Litani; Because user knows that candidate item 2 is translation results of candidate item 1, therefore select to knock in numeral 1, the upper screen of corresponding candidate item " Britney " can be completed.

It should be noted that, the present invention also may be used for the entirety translation of some long sentences or phrase, now, can pass through the mode of separator or participle, obtain required mixed defeated character string the long sentence inputted or phrase from user; Wherein, the mixed defeated character string in a long sentence can be one or more; Like this, the present invention can complete the input of meeting consumers' demand based on described one or more mixed defeated character string.

Due to the mixed defeated character string that described character input method inputs for user, more convenient user's input, make user's input process more natural and random, promote input efficiency and the accuracy rate of word/long sentence, and the different demands of user to transformation result can be met, such as, the mixed transformation result of defeated character string under source language and/or target language, or, described transformation result to the translation result of another kind of language, etc.

With reference to Fig. 9, show the structural drawing of a kind of input method system embodiment of the present invention, specifically can comprise:

Interface unit 901, for receiving the mixed defeated character string of user's input; Described mixed defeated character string can comprise pronunciation fragment and target language fragment;

Discrimination unit 902, for distinguishing mixed film trasport section dissimilar in mixed defeated character string;

Converting unit 903, for based on all dissimilar mixed film trasport sections, and obtains this mixed transformation result of defeated character string under target language according to preset processing rule;

Display unit 904, for showing transformation result;

Output unit 905, for according to user's instruction, exports the transformation result of specifying.

The input method system of Fig. 9 can be mainly used in translation, also can using the subsidiary function of interpretative function as input method, and namely this input method system also may comprise other modules of existing input method.That is, source language candidate item translation function that also can comprise existing input method of the input method system shown in Fig. 9 etc.Further, the present invention can be applied to various types of input method, as phonetic, five, GB, stroke etc., or the input method system Japanese, Korean etc. of other language.

The present embodiment for the mixed defeated character string of different situation, can design the structure of converting unit 903:

Situation 1,

When described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, the structure of described converting unit 903 can comprise:

Further, for transformation result acquisition module, the present embodiment can provide following design philosophy:

One, utilize pronunciation fragment and target language fragment in enunciative contact, the pronunciation unit sequence of described two fragments is integrated; Now, described transformation result acquisition module specifically can be in order to, based on the pronunciation unit sequence of pronunciation fragment under target language and the pronunciation unit sequence of target language fragment under target language, obtains this mixed transformation result of defeated character string under target language.

Two, utilize pronunciation fragment and the contact of target language fragment on character, the character string of described two fragments is connected; Based on this thought, following submodule can be designed in described transformation result acquisition module:

Splicing submodule, for obtained target language character string being connected with described target language fragment, obtaining this mixed transformation result of defeated character string under target language, and being shown.

Three, in the program, in described transformation result acquisition module, following submodule can be comprised:

Be appreciated that those skilled in the art can as required based on one or more of above-mentioned thought, carry out the structural design of transformation result acquisition module, the present invention is not limited this.

Situation 2,

When described dissimilar mixed film trasport section comprises the pronunciation fragment of the pronunciation fragment of target language and non-target language, described converting unit 903 can comprise:

Preferably, the translation result of transformation result also can also be presented to user by the input method system of Fig. 9, namely input method system now can also comprise: Language Translation unit, for the described mixed transformation result of defeated character string under target language is translated as another language words, and shown by display unit 904.

Preferably, when user carries out long sentence or phrase inputting, the input method system of Fig. 9 can also comprise: cutting unit, during for inputting long sentence or phrase as user, by the mode of separator or participle, the long sentence inputted from user or phrase, obtain required mixed defeated character string.

It should be noted that, in input method system of the present invention, concrete transfer process can complete in this locality, also can in the mode of telecommunication, there is provided translation result by one or several special translating servers, the present invention does not need to be limited this.Such as, the converting unit in aforementioned input method system can be positioned on remote server, and even, cutting unit also can be positioned on remote server, and user's input information is sent to remote server by local input method system, receives transformation result.

Each embodiment in this instructions all adopts the mode of going forward one by one to describe, and what each embodiment stressed is the difference with other embodiments, between each embodiment identical similar part mutually see.For device embodiment, due to itself and embodiment of the method basic simlarity, so description is fairly simple, relevant part illustrates see the part of embodiment of the method.

Above to a kind of method and apparatus carrying out character conversion based on mixed defeated character string provided by the present invention, and a kind of character input method and input method system, be described in detail, apply specific case herein to set forth principle of the present invention and embodiment, the explanation of above embodiment just understands method of the present invention and core concept thereof for helping; Meanwhile, for one of ordinary skill in the art, according to thought of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.

Claims

1. carry out a method for character conversion based on mixed defeated character string, it is characterized in that, comprising:

Obtain mixed defeated character string;

Distinguish mixed film trasport section dissimilar in mixed defeated character string, wherein, described mixed film trasport section comprises pronunciation fragment, and described pronunciation fragment is for determining the pronunciation unit sequence under target language;

Based on all dissimilar mixed film trasport sections, and obtain this mixed transformation result of defeated character string under target language according to preset processing rule.

2. the method for claim 1, it is characterized in that, described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, described based on all dissimilar mixed film trasport sections, and comprise according to step that preset processing rule obtains this mixed transformation result of defeated character string under target language:

Obtain the pronunciation unit sequence of pronunciation fragment under target language in described mixed defeated character string;

Based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, obtain this mixed transformation result of defeated character string under target language.

3. the method for claim 1, it is characterized in that, described dissimilar mixed film trasport section comprises the pronunciation fragment of target language and the pronunciation fragment of non-target language, described based on all dissimilar mixed film trasport sections, and comprise according to step that preset processing rule obtains this mixed transformation result of defeated character string under target language:

Based on the pronunciation unit sequence under target language, obtain this mixed transformation result of defeated character string under target language.

4. method as claimed in claim 2, is characterized in that, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising:

Obtain the pronunciation unit sequence of target language fragment under target language;

Based on the pronunciation unit sequence of pronunciation fragment under target language and the pronunciation unit sequence of target language fragment under target language, obtain this mixed transformation result of defeated character string under target language.

5. method as claimed in claim 2, is characterized in that, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising:

Based on the pronunciation unit sequence of pronunciation fragment under target language, obtain corresponding target language character string;

Obtained target language character string is connected with described target language fragment, obtains this mixed transformation result of defeated character string under target language.

6. method as claimed in claim 2, is characterized in that, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising:

Based on described target language fragment, in dictionary, matching inquiry obtains corresponding words set;

Utilize the pronunciation unit sequence of pronunciation fragment under target language, in described words set, search obtains occurrence, as this mixed transformation result of defeated character string under target language.

7. method as claimed in claim 2, is characterized in that,

Using the first fragment in described mixed defeated character string as pronunciation fragment, the second fragment, as target language fragment, obtains the transformation result under first object language;

8. method as claimed in claim 3, is characterized in that,

9. the method for claim 1, is characterized in that,

By information such as character types, upper and lower case letter, separator or state switching points, distinguish mixed film trasport section dissimilar in mixed defeated character string.

10. the method for claim 1, is characterized in that, also comprises:

The described mixed transformation result of defeated character string under target language is translated as another language words show.

11. the method for claim 1, is characterized in that, when user inputs long sentence or phrase, also comprise: by the mode of separator or participle, obtain required mixed defeated character string the long sentence inputted or phrase from user.

12. 1 kinds are carried out the device of character conversion based on mixed defeated character string, it is characterized in that, comprising:

Interface unit, for obtaining mixed defeated character string;

Discrimination unit, for distinguishing mixed film trasport section dissimilar in mixed defeated character string, wherein, described mixed film trasport section comprises pronunciation fragment, and described pronunciation fragment is for determining the pronunciation unit sequence under target language;

13. devices as claimed in claim 12, is characterized in that, described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, and described converting unit comprises:

14. devices as claimed in claim 12, is characterized in that, described dissimilar mixed film trasport section comprises the pronunciation fragment of target language and the pronunciation fragment of non-target language, and described converting unit comprises:

15. devices as claimed in claim 13, is characterized in that, described transformation result acquisition module specifically for:

16. devices as claimed in claim 13, it is characterized in that, described transformation result acquisition module comprises:

17. devices as claimed in claim 13, it is characterized in that, described transformation result acquisition module comprises:

18. devices as claimed in claim 12, is characterized in that, also comprise:

Language Translation unit, for being translated as another language words by the described mixed transformation result of defeated character string under target language.

19. devices as claimed in claim 12, is characterized in that, also comprise:

Cutting unit, during for inputting long sentence or phrase as user, by the mode of separator or participle, obtains required mixed defeated character string the long sentence inputted or phrase from user.

20. 1 kinds of character input methods, is characterized in that, comprising:

Receive the mixed defeated character string of user's input;

Based on all dissimilar mixed film trasport sections, and obtain this mixed transformation result of defeated character string under target language according to preset processing rule, and shown; According to user's instruction, export the transformation result of specifying.

21. methods as claimed in claim 20, it is characterized in that, described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, described based on all dissimilar mixed film trasport sections, and comprise according to step that preset processing rule obtains this mixed transformation result of defeated character string under target language:

22. methods as claimed in claim 20, it is characterized in that, described dissimilar mixed film trasport section comprises the pronunciation fragment of target language and the pronunciation fragment of non-target language, described based on all dissimilar mixed film trasport sections, and comprise according to step that preset processing rule obtains this mixed transformation result of defeated character string under target language:

23. methods as claimed in claim 21, is characterized in that, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising:

24. methods as claimed in claim 21, is characterized in that, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising:

25. methods as claimed in claim 21, is characterized in that, the described step obtaining transformation result based on the pronunciation unit sequence of pronunciation fragment under target language and target language fragment, comprising:

26. methods as claimed in claim 21, is characterized in that,

And using the second fragment in described mixed defeated character string as pronunciation fragment, the first fragment, as target fragment, obtains the transformation result under the second target language;

27. methods as claimed in claim 22, is characterized in that,

28. methods as claimed in claim 20, is characterized in that, also comprise:

29. methods as claimed in claim 20, is characterized in that, when user inputs long sentence or phrase, also comprise: by the mode of separator or participle, obtain required mixed defeated character string the long sentence inputted or phrase from user.

30. 1 kinds of input method systems, is characterized in that, comprising:

Display unit, for showing transformation result;

31. input method systems as claimed in claim 30, is characterized in that, described dissimilar mixed film trasport section comprises pronunciation fragment and target language fragment, and described converting unit comprises:

32. input method systems as claimed in claim 30, is characterized in that, described dissimilar mixed film trasport section comprises the pronunciation fragment of target language and the pronunciation fragment of non-target language, and described converting unit comprises:

33. input method systems as claimed in claim 31, is characterized in that, described transformation result acquisition module specifically for:

34. input method systems as claimed in claim 31, it is characterized in that, described transformation result acquisition module comprises:

35. input method systems as claimed in claim 31, it is characterized in that, described transformation result acquisition module comprises:

36. input method systems as claimed in claim 30, is characterized in that, also comprise:

37. input method systems as claimed in claim 30, is characterized in that, also comprise: